Behavioral variation in prey odor responses in northern pine snake neonates and adults

RESEARCH PAPER

Behavioral variation in prey odor responses in northern pinesnake neonates and adults

Kevin P. W. Smith • M. Rockwell Parker •

Walter F. Bien

Received: 14 October 2014 / Accepted: 23 February 2015

� Springer Basel 2015

Abstract Squamate reptiles (snakes, lizards, amphisbae-

nians) rely heavily on chemosensory cues to identify,

locate and choose between suitable prey items, but com-

paratively little research has focused on the chemical

ecology of threatened squamate species. Such knowledge

highlights ecologically important aspects of their survival.

Due to gape limitations, squamates often demonstrate on-

togenetic shifts in their diet where they consume larger

prey as they grow older and their gape size increases. This

shift enables squamates—especially snakes—to exploit

new resources in their environments, usually mammalian

prey. To test for ontogenetic variation in prey odor re-

sponses of a threatened snake species, the Northern pine

snake (Pituophis melanoleucus melanoleucus), we pre-

sented food-naıve neonates and food-experienced adults

with potential prey and non-prey animal scents and quan-

tified their behavioral responses. Our data indicate a strong

response to rodent scents from both neonates and adults.

Further, neonates showed more frequent investigative

probing and retreating behaviors from scented swabs and a

higher rate of tongue-flicking than adults. We also devel-

oped a new metric for measuring snake responses to prey

odor, a tongue-flick reaction score (TFRS), that incorpo-

rates investigative behaviors that may be unique to

constrictor-type snakes. The TFRS did not differ between

age classes and was highest when rodent odors were tested.

A canonical discriminant analysis confirmed the relation-

ship between TFRS behavioral components and tested

chemical signal reactions. Based on our data, P. me-

lanoleucus may fall into a category of snakes that exhibit

an ontogenetic telescope rather than a general ontogenetic

shift in their prey odor responses.

Keywords Prey odor � Behavior � Ontogeny �Tongue-flick � Squamata � Pituophis melanoleucus �Neonate

Introduction

Chemosensation is utilized by terrestrial vertebrates to

discriminate among a diverse array of environmental che-

mical signals. In snakes, the chemosensory system is used

for conspecific communication such as mate discovery and

selection (Cooper and Garstka 1987; LeMaster and Mason

2001; Schubert et al. 2008; Mason and Parker 2010; Shine

and Mason 2012), sibling recognition (Clark 2004), and

conspecific trailing to overwintering sites (Graves et al.

1991). Chemosensation-enabled heterospecific discrimina-

tion falls into two major categories: predator/threat

avoidance and prey detection. The avoidance of predator

chemical cues has been documented in reptiles, including

the Northern pine snake, Pituophis melanoleucus (Burger

et al. 1991), kingsnake, Lampropeltis getula (Weldon and

Schell 1984), and desert iguana, Dipsosaurus dorsalis

Taxa Class: Reptilia—Order: Squamata—Family: Colubridae—Genus: Pituophis—Species: Melanoleucus—Subspecies:Melanoleucus.

Handling Editor: Michael Heethoff.

K. P. W. Smith (&) � W. F. Bien

Biodiversity, Earth, and Environmental Science Department,

Drexel University, Philadelphia, PA 19104, USA

e-mail: [email protected]

M. R. Parker

Monell Chemical Senses Center, Philadelphia, PA 19104, USA

M. R. Parker

Department of Biology, Washington and Lee University,

Lexington, VA 24450, USA

Chemoecology

DOI 10.1007/s00049-015-0193-6 CHEMOECOLOGY

123

(Bealor and Krekorian 2006). The importance of

chemosensation in detecting and discriminating among

prey has been the subject of much study (Burghardt and

Hess 1968; Halpern and Frumin 1979; Amo et al. 2004;

Stark et al. 2011). Chemosensation is associated with a

suite of observable behaviors that lend themselves to

quantifying individual interests in available chemical cues

and enables evolutionary and ontogenetic inquiry (Cooper

and Burghardt 1990).

Tongue-flicking is one of the most identifiable behaviors

linked with the vomeronasal system of squamates (Gove

and Burghardt 1975; Mason 1992). The tongue protrudes

from the mouth to collect mostly non-volatile compounds

from the surrounding surfaces and deliver them to the

vomeronasal organ (VNO) for processing (Schwenk 1993).

Airborne chemical cues as well as superficial deposits can

aid in scent source location in multiple reptiles (Waters

1993; Cooper and Perez-Mellado 2001; Parker and Kar-

dong 2005). The presence of a distinct VNO–oral junction

from the main olfactory system in squamates allows for the

rapid collection and separate assessment of chemical sig-

nals; such organization is absent in Crocodylia and

Sphenodontia (Kaas 2009). The loss or blockage of the

VNO duct reduces or prevents discriminatory responses to

chemical cues (Reformato et al. 1983; Stark et al. 2011).

Though the VNS is the central pathway for detecting and

responding to prey odors, many animal species use nasal

olfaction and/or gustation for assessing food cues as well

(e.g., Terrick et al. 1995; Cooper and Perez-Mellado 2001;

Saviola et al. 2012; Lopez et al. 2014).

The anatomical restrictions of a gape-limited predator

(e.g., snakes) will affect the niche the predator fills and can

shape prey body size as well. Body size increase is a

common and effective adaptation of prey to limit their

susceptibility to gape-limited predators (perch: Persson

et al. 1996; salamanders: Urban 2007, 2008). The coevo-

lutionary relationship between predator gape and prey size

not only affects prey community composition but also

predator diet selection over time. If the size of prey com-

pared to predator gape dictates the possibility of

consumption, then predator gape size affects prey selection

(Hampton 2014).

Ontogenetic shifts in diet preference are common in

gape-limited predators (Hampton and Moon 2013). Prey

availability increases with larger gape dimensions (Slip and

Shine 1988). Snakes are a particularly insightful group of

reptiles for studying the effects of gape limitation on prey

preference (Greene 1983; King 2002), especially the rela-

tionship between ontogeny and prey odor responses. Many

snake species have prey odor responses that are contingent

on a population’s habitat and can be plastic (e.g., garter

snakes: Burghardt 1993; Burghardt et al. 2000; pygmy

rattlesnake: Bevelander et al. 2006; striped crayfish snake:

Waters and Burghardt 2013). Other species display onto-

genetic shifts in diet and prey odor responses but lack

plasticity within life stages (Gove and Burghardt 1975;

Dunbar 1979; Shepard et al. 2004; Saviola et al. 2012).

Finally, few snakes are specialists that display no plasticity

in prey odor responses (e.g., queen snakes: Jackrel and

Reinert 2011; eastern hog-nose: Cooper and Secor 2007).

However, exposure to prey odors early in life can induce a

chemosensory preference to such odors in many snake

species, demonstrating that learning occurs in many taxa

(Loop 1970; Burghardt and Krause 1999; Aubret et al.

2006).

Our research focused on the putative prey cue responses

of a threatened species, the Northern pine snake (Pituophis

melanoleucus melanoleucus), both early in life and in

adulthood. As adults, Northern pine snakes, Pituophis

melanoleucus, have a variety of potential prey in their

natural environments (Diller and Wallace 1996). The prey

response of neonates to a suite of potential prey items has

yet to be investigated. Rodents encompass 70–93 % of the

typical adult Pituophis diet, with a smaller percentage be-

ing composed of birds, lagomorphs, and eggs (P.

melanoleucus: Diller and Wallace 1996; P. cantifer: Ro-

drıguez-Robles 2002). The cryptic and semi-fossorial

nature of this species makes in situ observations of adult

and especially neonates difficult to obtain. Natricine snakes

and vipers have a wide spectrum of degrees of plasticity in

their diets, potentially influenced by their gape limitation.

Investigating an additional colubrid species, P. me-

lanoleucus, will further explain the variations in prey

choice ontogeny and the role of neonates in their

ecosystem.

Methods

Study animals

We collected Northern pine snake, Pituophis melanoleucus

melanoleucus, neonates (n = 19; mean snout–vent

length = 42.7 ± 2.8 cm; mean mass = 37.9 ± 4.8 g)

from nests monitored in Franklin Parker Preserve, in

Chatsworth, New Jersey. Nests were corralled with 1 m silt

fencing and all neonates hatched naturally in subterranean

nests excavated by their mothers. We collected individuals

in one-way box traps that we checked daily and then

housed them in tanks according to nest of origin after their

first ecdysis (shed). No neonates tested showed signs of

recent meals (e.g., bolus) and were collected soon after the

first shed; thus they were food naıve. Hatchling pine snakes

will not take food before their first shed (Burger et al.

1987). Tests took place in September within 2 weeks of

hatching.

K. P. W. Smith et al.

123

We collected adult Northern pine snakes (n = 13; mean

SVL = 127.4 ± 10.9 cm; mean mass = 728.6 ± 250.6 g)

at Warren Grove Gunnery Range, in Warren Grove, New

Jersey. Known hibernacula were corralled with 1.2 m

hardwire cloth (0.64 cm mesh gauge) and individuals were

captured in one-way box traps. We performed tests im-

mediately after egress in early spring and no adults showed

any signs of recent meals (i.e., bolus present). All test

subjects were allowed water ad libitum.

Swab preparation

Neonates and adults were exposed to seven odors presented

on swabs: blank (hexane only), water (hexane plus water),

white-footed mouse (Peromyscus leucopus), fence lizard

(Sceloporus undulatus), Fowler’s toad (Bufo fowleri),

grasshopper (Oedipodinea), and cricket (Gryllidae). These

animals represent potential prey and non-prey items within

the gape limitations of a neonate pine snake. We captured

prey animals in the wild and swabbed with a hexane-dip-

ped cotton-tipped wooden applicator (Dynarex) to collect

lipid-based scents. Solvent extraction of vomodors results

in the isolation of relevant chemosensory stimuli that can

be used in behavioral assays (e.g., Graves and Halpern

1988; Weldon and Schell 1984; summarized in Mason

1992; Mason and Parker 2010). Thus, we chose to extract

odors using hexane to acquire chemical stimuli. Further,

hexane only extracts lipophilic and not aqueous prey odors,

and the majority of vomodors that squamate reptiles re-

spond to are lipophilic (Lopez et al. 2006; Mason 1992).

Wild-caught vertebrates were swabbed and released alive

unharmed after trials. Insects were sacrificed for sampling.

Rodents for adult tests were supplied by Monell Chemical

Senses Center and swabbed in the same procedure as wild-

caught rodents.

After swabbing, the applicators were allowed to dry to

ensure the hexane evaporated so as to avoid a behavioral

reaction to the solvent. Dried swabs were stored in plastic

bags and separated by scent then frozen overnight to limit

the loss of potentially bioactive molecules. Swabs were

allowed to reach ambient temperature (24–29 �C) at the

time of the trials.

Trial procedure

Chemosensory tests were conducted in sterilized 38-l glass

aquaria covered on the outside with opaque paper to

minimize visual stimuli. All tests were performed between

24–29 �C and 1100–1800 h (active temperature and time

of day for this diurnal species (Burger and Zappalorti

1992). Each snake was tested once through the battery of

scents. Animals were allowed to acclimate for 10 min in a

testing aquarium before and after the randomized

presentation of each of the seven swabs until all swabs had

been tested. Swabs were presented 2 cm from the snout of

the snake for 60 s, recording the rate of tongue-flicks (RTF,

tongue-flicks per minute), number of snout rubs, how long

it took for a snake to execute a snout rub [snout rub latency

(s)], number of C90� retreats where the animal turned

away from the swab (retreats), and retreat latency (s).

Score calculation

Pituophis melanoleucus does not display the same readi-

ness to strike swabs as observed in garter snakes

(Burghardt 1969, 1970) or lizards (Cooper et al. 1990;

Garrett and Card 1993). However, test subjects did display

a snout rubbing behavior indicative of chemosensory in-

vestigation as seen in other snake species during behavioral

trials (e.g., Scudder et al. 1980; Jackrel and Reinert 2011).

They also displayed retreating behaviors, suggesting lack

of interest or an adverse reaction. As a result of these

different behaviors, we considered that the traditional

tongue-flick attack score (TFAS) developed by Burghardt

(1969, 1970) would not fully describe the reaction of pine

snakes to chemical stimuli presented on swabs. Thus, the

projected tongue-flick rate (PTFR), still based on a pro-

jection contingent on attack, would also not apply to pine

snakes (Arnold 1978; Halpern and Frumin 1979). Thus, we

modified Burghardt’s original TFAS (1970) as follows. The

TFAS was calculated as:

TFAS ¼ Tongue�flicks þ Test length� Attack latencyð Þ

We included positive (investigative) behaviors (e.g.,

snout rubs: Jackrel and Reinert 2011) as well as negative

(disinterested) behaviors (e.g., retreating) to result in a

tongue-flick reaction score (TFRS):

TFRS ¼ Tongue�flicks

þ Snout rubsþ Test length� Rub latencyð Þ½ �� Retreatsþ Test length� Retreat latencyð Þ½ �

Representing

TFRS ¼ Tongue�flicks þ Positive behaviorsð Þ� Negative behaviorsð Þ

This score controls for the level of interest as Burghardt’s

1970 TFAS does while also including a disinterest com-

ponent to express active movement away from the stimuli.

Statistical analysis

Two-way repeated measures ANOVAs were used to test

for age effects and swab type effects in all behaviors. For

all significantly affected behaviors, as well as compiled

groups of age-independent behaviors, we used one-way

Behavioral variation in prey odor responses in northern pine snake neonates and adults

123

repeated measures ANOVAs. Reported degrees of freedom

reflect Greenhouse–Geisser correction for sphericity

(Table 1). Two-tailed t-tests were used for all pairwise

comparisons. All alphas set to a = 0.05. All tests and

graphs were produced with SPSS 22 (IBM Corp. Released

2013).

For the purpose of data presentation, swab types were

abbreviated as follows: blank (BL), water (WA), rodent

(RO), fence lizard (FL), toad (TD), grasshopper (GH), and

cricket (CR).

Canonical discriminant analysis

A canonical discriminant analysis (CDA) was used to de-

termine the efficacy of categorizing scent-related

behavioral responses with the monitored behaviors (RTF,

snout rubs, contact latency, retreats, and retreat latency).

Canonical discriminant analysis uses factors to form a

discriminant function, which is then tested for its dis-

criminant ability to group individual responses into

categories (Kramer et al. 2009). This test was performed to

confirm whether the behavioral factors are associated with

scent signal reactions. The discriminant functions were

compiled based on the Wilks’ Lambda scores of individual

factors and then tested with a Chi square for its dis-

criminating ability in SPSS 22 (IBM Corp. Released 2013).

Results

Rate of tongue-flick (RTF)

Two-way RM ANOVA shows that there was a significant

effect between age and RTF (F = 5.768, p = 0.033). Swab

type, however, did not have an effect on RTF (F = 0.743,

p = 0.511). The age effect was due to neonates having

higher RTFs than adults for fence lizard swabs (q = 3.05,

p = 0.031) and cricket swabs (q = 3.21, p = 0.023).

Neonate RTF for grasshopper swabs was marginally sig-

nificantly different from that of adults (q = 2.76,

p = 0.051). No other within-swab comparisons were sig-

nificantly different (p [ 0.1) (Fig. 1).

Snout rubs

Age (F = 6.165, p = 0.029) and swab type (F = 5.605,

p = 0.021) had significant effects on the number of snout

rubs (two-way RM ANOVA). There was no interaction

between age and swab type (F = 1.053, p = 0.376). The

age effect was due to neonates having a marginally higher

number of snout rubs for fence lizards than did adults

(q = 2.125, p = 0.055). When age group data (neonates

and adults) were pooled, the number of rodent snout rubs

Table 1 Repeated-measures ANOVA of swab odor responses within and across age classes

Adult (n = 13) Neonate (n = 19) Adult (n = 13) vs. Neonate (n = 19)

df F p value df F p value df F p value

RTF 4.209 0.628 0.653 3.877 1.649 0.173 1 5.768 0.033

Snout rubs 1.655 3.78 0.048 2.031 5.768 0.006 1 6.165 0.029

Contact latency – – – – – – 1 3.436 0.089

Retreats 2.698 2.677 0.069 3.487 1.388 0.252 1 10.467 0.007

Retreat latency 2.796 2.239 0.106 5.606 1.209 0.314 1 9.416 0.010

TFRS – – – – – – 1 1.094 0.316

Two-way RM ANOVA compared age as a factor in response for Adult vs. Neonate. One-way RM ANOVA compares the effect of swab type

within age classes if age was determined to be a factor. Significant results are bold. p values associated with F-statistics are corrected via

Greenhouse–Geisser

Fig. 1 Mean rate of tongue-flicking (tongue-flicks/min; RTF)

(±SEM) of adult (gray bars) and neonate (white bars) pine snakes

when exposed to swabs containing different prey odors. Neonates had

higher average RTFs than adults (p = 0.033), and neonate RTF was

higher than adults for CR and FL swabs. Asterisks represent

significant differences between age classes (p \ 0.05)

K. P. W. Smith et al.

123

was significantly higher than all scents (BL, q = 4.20,

p \ 0.001; WA, q = 3.98, p \ 0.001, FL, q = 3.36,

p = 0.002; TD, q = 3.00, p = 0.005; GH, q = 2.98,

p = 0.006; CR, q = 3.06, p = 0.005). There tended to be

fewer blank snout rubs than for fence lizard (q = 1.856,

p = 0.073) and significantly fewer than for other animal

scents (TD, q = 2.35, p = 0.025; GH, q = 2.35,

p = 0.025; CR, q = 2.56, p = 0.016). There were sig-

nificantly fewer water snout rubs than for all animal scents

(FL, q = 2.10, p = 0.044; TD, q = 2.35, p = 0.025; GH,

q = 2.52, p = 0.017; CR, q = 2.37, p = 0.024). Snout

rubs for blank and water did not differ significantly

(p [ 0.1) (Fig. 2).

Adults (F = 3.78, p = 0.048) and neonates (F = 5.768,

p = 0.006) had a significant difference between snout rubs

and scents. In adults, snout rubs for rodent scent were

marginally greater than for cricket scent (q = 1.78,

p = 0.1) and grasshopper scent (q = 2.10, p = 0.057), and

significantly greater than for blank (q = 2.56, p = 0.025),

water (q = 2.56, p = 0.025) and fence lizard scent

(q = 2.56, p = 0.025). All other adult pairwise compar-

isons between scents were not significantly different

(p [ 0.1).

In neonates, snout rubs for rodent scent were sig-

nificantly higher than for all other scents (BL, q = 3.37,

p = 0.003; WA, q = 3.13, p = 0.006, FL, q = 2.41,

p = 0.027; TD, q = 2.50, p = 0.022; GH, q = 2.24,

p = 0.038; CR, q = 2.43, p = 0.026). Snout rubs for blank

were marginally lower in number than for toad (q = 1.92,

p = 0.07), fence lizard (q = 1.91, p = 0.072), and

grasshopper scents (q = 1.96, p = 0.066) and significantly

lower in number than cricket scent (q = 2.39, p = 0.028).

Snout rubs for water were marginally fewer than for toad

scent (q = 1.92, p = 0.07) and significantly fewer than for

the rest of the animal scents (FL, q = 2.19, p = 0.042;

GH, q = 2.14, p = 0.047; CR, q = 2.17, p = 0.044). All

other pairwise comparisons between scents were not sig-

nificantly different (p [ 0.1).

Contact latency

Age class had a marginally significant effect (two-way RM

ANOVA) on contact latency (F = 3.436, p = 0.089).

Swab type also had a significant effect on contact latency

(F = 5.48, p \ 0.001). We combined adults and neonates

due to the marginal age effect. When neonates and adults

were grouped together without age as a factor, there were

significant differences among contact latencies

(df = 4.173, F = 5.728, p \ 0.0001). Contact latency for

rodent swabs was significantly less than all other swab

types (BL, q = 6.78, p \ 0.001; WA, q = 7.15,

p \ 0.001, FL, q = 4.98, p = 0.004; TD, q = 4.06,

p = 0.004; GH, q = 4.52, p = 0.004; CR, q = 4.69,

p = 0.005). All other pairwise comparisons between scents

were not significantly different (p [ 0.1) (Fig. 3).

Retreats

There was a significant effect (two-way RM ANOVA)

between age and retreat amount (F = 10.467, p = 0.007),

but swab type was not significant (F = 0.936, p = 0.423).

Fig. 2 Mean number of snout rubs (±SEM) of adult (gray bars) and

neonate (white bars) pine snakes when exposed to swabs containing

different prey odors. Neonates had a higher number of snout rubs than

adults (p = 0.029). Capital letters represent adult significant differ-

ences between swab types (p \ 0.05). Asterisk represents neonate

significance difference from all other scents (p \ 0.05)

Fig. 3 Mean contact latency (s) (±SEM) of adult and neonate pine

snakes when exposed to swabs containing different prey odors.

Contact latency for rodent scent was shorter than all other scents.

Asterisk represents significant differences between contact latency in

response to scents (p \ 0.05)

Behavioral variation in prey odor responses in northern pine snake neonates and adults

123

The age effect was due to neonates having higher number

of retreats than adults for blank (q = 2.27, p = 0.042),

water (q = 2.96, p = 0.12), rodent (q = 2.65, p = 0.021),

fence lizard (q = 3.08, p = 0.01), and cricket (q = 2.29,

p = 0.041). Adults showed marginal differences in number

of retreats between swab types (F = 2.677, p = 0.069),

but neonates did not (F = 1.388, p = 0.252) (Fig. 4).

Retreat latency

There was a significant effect (two-way RM ANOVA)

between age and retreat latency (F = 9.416, p = 0.010),

and swab type had a marginal effect (F = 2.189,

p = 0.051). The age effect was due to neonates having

marginally shorter retreat latencies than adults for blank

(q = 1.93, p = 0.077) and significantly shorter retreat la-

tencies for water (q = 3.11, p = 0.009), rodent (q = 2.67,

p = 0.021), fence lizard (q = 3.02, p = 0.011) and cricket

(q = 2.40, p = 0.034). There were no significant effects

from swab type in either adults (F = 2.239, p = 0.106) or

neonates (F = 1.209, p = 0.314) (Fig. 5).

Tongue-flick reaction score tests

Age class did not have a significant effect (two-way RM

ANOVA) on TFRS (F = 1.094, p = 0.316), but swab type

did have a significant effect (F = 6.04, p \ 0.001). When

all snakes were grouped together without age as a factor,

there were significant differences among TFRS

(df = 4.741, F = 6.415, p \ 0.001). The TFRS for rodent

scent was significantly higher than for all other scents (BL,

q = 7.79, p \ 0.001; WA, q = 5.92, p \ 0.001, FL,

q = 5.90, p \ 0.001; TD, q = 5.379, p \ 0.001; GH,

q = 6.22, p \ 0.001; CR, q = 4.46, p = 0.002). There

were no other differences in TFRS between swab types

(p [ 0.1) (Fig. 6).

Field observation

In September 2012, we collected an untested neonate from

one of the nest sites at Franklin Parker Preserve 2 weeks

after release of the animals used in this study. The neonate

Fig. 4 Mean number of retreats (±SEM) of adult (gray bars) and

neonate (white bars) pine snakes when exposed to swabs containing

different prey odors. Neonates had a higher number of retreats than

adults (p = 0.007). Asterisks represent significant differences be-

tween age classes (p \ 0.05)

Fig. 5 Mean retreat latency (s) (±SEM) of adult (gray bars) and

neonate (white bars) pine snakes when exposed to swabs containing

different prey odors. Neonates had a higher number of retreats than

adults (p = 0.007). Asterisks represent significant differences be-

tween age classes (p \ 0.05)

Fig. 6 Mean TFRS (± SEM) of adult and neonate pine snakes when

exposed to swabs containing different prey odors. Pine snakes scored

a higher TFRS for rodents than any other scent. Asterisks represent

significantly different scores in response to scents (p \ 0.05)

K. P. W. Smith et al.

123

had a large bolus. After 2 days of captivity alone, the

neonate regurgitated the bolus. The presence of dark

hairs [1 cm long indicate adult rodent prey.

Canonical discriminant analysis

Among the seven scent trials, the combination of five

discriminant functions including all behavioral factors had

a marginally significant discriminating ability in adult tests

(Wilks’ Lambda = 0.604; df = 30; p = 0.069) and sig-

nificant ability in neonate tests (WL = 0.628; df = 30;

p = 0.001).

We repeated the analysis after grouping the seven scent

trials into three groups (Control: blank, water; Other: fence

lizard, toad, grasshopper, cricket; Rodent). When dis-

criminating between these three scent groups, the

combination of two discriminant functions in both adults

(WL = 0.737; df = 10; p = 0.003) and neonates

(WL = 0.717; df = 10; p \ 0.001) had significant dis-

criminating ability. In the grouped trials, the first function

for each group contained 85 % of the variation (Table 2).

Discussion

Neonate pine snakes differentiated between various lipid-

based chemical scents extracted from live animals as the

adults did. However, we were also able to quantify dif-

ferent interest levels between presented odor cues and

between neonates and adults based on tested behaviors.

Hexane swabs were effective at extracting lipophilic

biologically relevant signals, a method consistent with that

reported in the literature (Cooper and Garstka 1987; Mason

1992; Bealor and Krekorian 2006). The signals were strong

enough to elicit measurable behavioral responses. The re-

lationships between the behaviors contributing to the TFRS

and the behavioral reactions to scent signals were sup-

ported through a canonical discriminate analysis (Kramer

et al. 2009). These relationships were evident when di-

viding the scent groups into three sub-groups.

Adult and neonate pine snakes showed significantly

variable responses across multiple chemosensory behav-

iors. Neonates had a higher RTF and rubbed swabs more

across all scents and retreated more often and more quickly

than adults. Such behaviors indicate heightened investiga-

tion/exploration on the part of neonates. Given their

presumed naıve food exposure and comparative lack of

experience versus adults, these extra investigatory behav-

iors may suggest initial learning. In at least one species of

snake, Coelognathus helena, neonates develop prey-han-

dling techniques over repetitive exposure (Mehta 2008).

Further, venomous snake species develop abilities to meter

out venom dosage as well (Hayes 1995), and in constric-

tors, extra probing and physical contact can initiate

constriction (Greene and Burghardt 1978) or relate to in-

vestigative probing for prey (Jackrel and Reinert 2011).

Neonate pine snakes have a significant reaction to mouse

scent if previously fed mice (Burger 1991), and it appears

that neonates investigated the swabs as part of an odor-

learning process. Our study supports strong response to

rodent odor reported by Burger (1991) and also shows an

overall lack of a significant response to other odors rep-

resenting potential prey items.

Our TFRS data suggest a high level of interest in rodent

prey at both the adult and neonate stage. This was un-

surprising for adults based on previous work (de Queiroz

1984). The neonate interest in rodents we observed is

further supported by the collection of a regurgitated

sample from a neonate that contained adult rodent hairs.

Neonate pine snakes showed low-level responses to the

majority of prey odors presented, though these cues came

from readily available species that are small and more

easily captured than rodents. Adult Northern pine snakes

primarily feed on rodents, with birds, eggs and lago-

morphs represented in smaller proportions of their diets

(P. melanoleucus: Diller and Wallace 1996; P. cantifer:

Table 2 Canonical discriminant function coefficients for behavioral factors contributing to two functions

Behavior factors Functions

Adults Neonate

1 2 1 2

Snout rubs 0.80 0.37 0.96 0.83

Contact latency -0.21 0.72 -0.02 0.98

Tongue-flick 0.04 0.67 0.15 -0.47

Retreat 0.22 -0.91 -0.14 1.10

Retreat latency 0.46 -0.19 0.46 1.00

% Of variance 85.4 14.6 84.5 15.5

The two functions for each age class were used in unison to collate behavioral variation in reaction to three scent groups: control, other, and

rodent. The first function for each age group contains roughly 85 % of variation

Behavioral variation in prey odor responses in northern pine snake neonates and adults

123

Rodrıguez-Robles 2002;). The overlap of rodent prey

between neonates and adults, however, suggests an ‘‘on-

togenetic telescope’’—rather than a shift—in prey

selection (King 2002). As neonate pine snakes grow, they

may include other auxiliary items; however, they are non-

plastic in their predation of rodents.

Food-naıve hatchling pine snakes may be spending more

effort investigating potentially dangerous prey (e.g., ro-

dents) rather than less threatening prey readily available in

their habitat (e.g., insects). The danger of engaging live

rodents is a strong selective force for snakes and has been

proposed as a driver of the evolution of strike-and-release

tactics in venomous snakes that gave rise to a complex

behavior pattern called strike-induced chemosensory

searching (Chiszar et al. 1976; Kardong 1986; Clark 2006).

The relationship between predatory behavior and potential

injury is also affected by the size of prey. Rattlesnakes

typically strike-and-release prey but will strike and hold

prey if the prey are on the smaller end of the potential size

range (Radcliffe et al. 1980). Pine snakes, however, are

constrictors, and therefore do not release prey upon con-

tact. Prey-handling proficiency in some constrictors can

change significantly with age and experience (Mori 1994).

This makes the successful consumption of an adult rodent

by a 2-week-old neonate that much more dangerous, and it

may be that neonates are adept constrictors, possess novel

behaviors for prey handling and/or primarily consume

smaller, more easily handled rodent prey.

The preservation of prey selection from neonate to adult

in a gape-limited predator presents distinct advantages and

disadvantages. Taking large prey as a neonate can be po-

tentially dangerous, however, consumption of a large item

at such a small predator:prey size ratio could mean a high

net energy gain (Forsman 1996; Troost et al. 2008). Prey

handling as a full-grown adult would be easier (de Queiroz

1984), though more prey items would need to be consumed

to offset the diminishing returns. Pituophis melanoleucus

hatch at *32 g and *47 cm SVL (Burger et al. 1987). At

an average adult mass of near 800 g, the neonates body

mass increases around 259 by adulthood (Gerald et al.

2006). Our data suggest that rather than an ontogenetic

shift in diet, there may be an ontogenetic shift in foraging

behavior or frequency that compensates for different needs

at different life stages (Lind and Welsh 1994).

Acknowledgments This study was conducted under NJDEP state

permits (Permit No. SC 2012-085, SC 2013-085, SC 2014-085) and

Drexel University IACUC (18924 and 20129). Thank you to the New

Jersey Air National Guard and New Jersey Conservation Foundation

for access to research sites. A special thanks to the Laboratory of

Pinelands Research at Drexel University for financial support. KPWS

would especially like to thank Dr. Emile DiVito, Raffaella Marano,

Emily Ostrow, Kathryn Bendon and members of the Laboratory of

Pinelands Research for assistance with field work and experiments,

Dr. James R. Spotila for aid in manuscript revisions, and Kevin

Redding at Monell Chemical Senses Center for supplying rodent

scents for tests. The authors would also like to thank an anonymous

reviewer for comments that significantly improved the manuscript.

References

Amo L, Lopez P, Martın J (2004) Chemosensory recognition of its

lizard prey by the ambush smooth snake, Coronella austriaca.

J Herpetol 38(3):451–454

Arnold SJ (1978) Some effects of early experience on feeding

responses in the common garter snake, Thamnophis sirtalis.

Anim Behav 26:455–462

Aubret F, Burghardt GM, Maumelat S, Bonnet X, Bradshaw D (2006)

Feeding preferences in 2 disjunct populations of tiger snakes,

Notechis scutatus (Elapidae). Behav Ecol 17(5):716–725

Bealor MT, Krekorian CO (2006) Chemosensory response of desert

iguanas (Dipsosaurus dorsalis) to skin lipids from a lizard-eating

snake (Lampropeltis getula californiae). Ethology 112:503–509

Bevelander G, Smith T, Kardong K (2006) Microhabitat and prey

odor selection in the foraging pigmy rattlesnake. Herpetologica

62(1):47–55

Burger J (1991) Response to prey chemical cues by hatchling pine

snakes (Pituophis melanoleucus): effects of incubation tem-

perature and experience. J Chem Ecol 17:1069–1078

Burger J, Zappalorti RT (1992) Philopatry and nesting phenology of

pine snakes Pituophis melanoleucus in the New Jersey Pine

Barrens. Behav Ecol Sociobiol 30(5):331–336

Burger J, Zappalorti R, Gochfeld M (1987) Developmental effects of

incubation temperature on hatchling pine snakes Pituophis

melanoleucus. Comp Biochem 87:727–732

Burger J, Boarman W, Kurzava L, Gochfeld M (1991) Effect of

experience with pine (Pituophis melanoleucus) and king (Lam-

propeltis getulus) snake odors on Y-maze behavior of pine snake

hatchlings. J Chem Ecol 17:79–87

Burghardt G (1969) Comparative prey-attack studies in newborn

snakes of the genus Thamnophis. Behaviour 33:77–114

Burghardt G (1970) Intraspecific geographical variation in chemical

food cue preferences of newborn garter snakes. Behaviour

36:246–257

Burghardt G (1993) The comparative imperative: genetics and

ontogeny of chemoreceptive prey response in Natricine snakes.

Brain Behav Evol 41(138–1):46

Burghardt GM, Hess EH (1968) Factors influencing the chemical

release of prey attack in newborn snakes. J Comp Physiol

Psychol 66(2):289–295

Burghardt G, Krause M (1999) Plasticity of foraging behavior in

garter snakes (Thamnophis sirtalis) reared on different diets.

J Comp Psychol 113(3):277–285

Burghardt G, Layne D, Konigsberg L (2000) The genetics of dietary

experience in a restricted natural population. Psychol Sci

11:69–72

Chiszar D, Scudder K, Knight L (1976) Rate of tongue-flicking by

garter snakes (Thamnophis radix haydeni) and rattlesnakes

(Crotalus v. viridis, Sistrurus catenatus tergeminus, and S. c.

edwardsi) during prolonged exposure to food Odors. Behav. Biol

283(5233):273–283

Clark RW (2004) Kin recognition in rattlesnakes. Proc Biol Sci

271:243–245

Clark RW (2006) Post-strike behavior of timber rattlesnakes (Cro-

talus horridus) during natural predation events. Ethology

112:1089–1094

Cooper W, Burghardt G (1990) A comparative analysis of scoring

methods for chemical discrimination of prey by squamate

reptiles. J Chem Ecol 16(1):45–65

K. P. W. Smith et al.

123

Cooper W, Garstka W (1987) Lingual responses to chemical fractions

of urodaeal glandular pheromone of the skink Eumeces laticeps.

J Exp Zool 242:249–253

Cooper WE, Perez-Mellado V (2001) Chemosensory responses to

sugar and fat by the omnivorous lizard Gallotia caesaris: with

behavioral evidence suggesting a role for gustation. Physiol

Behav 73(4):509–516

Cooper W, Secor S (2007) Strong response to anuran chemical cues

by an extreme dietary specialist, the eastern hog-nosed snake

(Heterodon platirhinos). Can J Zool 85(5):619–625

Cooper WE, Buth DG, Vitt LJ (1990) Prey odor discrimination by

ingestively naive coachwhip snakes (Masticophis flagellum).

Chemoecology 1:86–91

de Queiroz A (1984) Effects of prey type on the prey-handling

behavior of the bullsnake, Pituophis melanoleucus. J Herpetol

18:333–336

Diller L, Wallace R (1996) Comparative ecology of two snake species

(Crotalus viridis and Pituophis melanoleucus) in Southwestern

Idaho. Herpetologica 52:343–360

Dunbar GL (1979) Effects of early feeding experience on chemical

preference of the Northern water snake, Natrix s. sipedon.

J Herpetol 13:165–169

Forsman A (1996) Body size and net energy gain in gape-limited

predators: a model. J Herpetol 30:307–319

Garrett C, Card W (1993) Chemical discrimination of prey by naive

neonate Gould’s monitors Varanus gouldii. J Chem Ecol

19:2599–2604

Gerald G, Bailey M, Holmes J (2006) Movements and activity range

sizes of Northern pine snakes (Pituophis melanoleucus me-

lanoleucus) in Middle Tennessee. J Herpetol 40:503–510

Gove D, Burghardt G (1975) Responses of ecologically dissimilar

populations of the water snake Natrix s. sipedon to chemical cues

from prey. J Chem Ecol 1:25–40

Graves BM, Halpern M (1988) Neonate plains garter snakes

(Thamnophis radix) are attracted to conspecific skin extracts.

J Comp Psychol 102:251–253

Graves B, Halpern M, Friesen J (1991) Snake aggregation

pheromones: source and chemosensory mediation in western

ribbon snakes (Thamnophis proximus). J Comp Psychol

105:140–144

Greene H (1983) Dietary correlates of the origin and radiation of

snakes. Am Zool 23:431–441

Greene H, Burghardt G (1978) Behavior and phylogeny: constriction

in ancient and modern snakes. Science 200(4337):74–77

Halpern M, Frumin N (1979) Roles of the vomeronasal and olfactory

systems in prey attack and feeding in adult garter snakes. Physiol

Behav 22:1183–1189

Hampton PM (2014) Allometry of skull morphology, gape size and

ingestion performance in the banded watersnake (Nerodia

fasciata) feeding on two types of prey. J Exp Biol 217(Pt

3):472–478

Hampton PM, Moon BR (2013) Gape size, its morphological basis,

and the validity of gape indices in western diamond-backed

rattlesnakes (Crotalus atrox). J Morphol 274:194–202

IBM Corp. Released (2013) IBM SPSS Statistics for Windows,

Version 22.0. IBM Corp., Armonk, NY. http://www-01.ibm.

com/support/docview.wss?uid=swg21476197

Hayes W (1995) Venom metering by juvenile prairie rattlesnakes,

Crotalus v. viridis: effects of prey size and experience. Anim

Behav 50:33–40

Jackrel S, Reinert H (2011) Behavioral responses of a dietary

specialist, the queen snake (Regina septemvittata), to potential

chemoattractants released by its prey. J Herpetol 45:272–276

Kaas JH (2009) Evolutionary neuroscience. Academic Press, Oxford,

pp 428–431

Kardong KV (1986) Predatory strike behavior of the rattlesnake,Crotalus viridis oreganus. J Comp Psychol 100:304–314

King R (2002) Predicted and observed maximum prey size–snake size

allometry. Funct Ecol 16:766–772

Kramer M, Weldon PJ, Carroll JF (2009) Composite scores for

concurrent behaviours constructed using canonical discriminant

analysis. Anim Behav 77:763–768

LeMaster MP, Mason RT (2001) Evidence for a female sex pheromone

mediating male trailing behavior in the red-sided garter snake,

Thamnophis sirtalis parietalis. Chemoecology 11:149–152

Lind A, Welsh HH (1994) Ontogenetic changes in foraging behaviour

and habitat use by the Oregon garter snake, Thamnophis atratus

hydrophilus. Anim Behav 48(6):1261–1273

Loop M (1970) The effects of feeding experience on the response to

prey-object extracts in rat snakes. Psychon Sci 21:189–190

Lopez P, Amo L, Martın J (2006) Reliable signaling by chemical cues

of male traits and health state in male lizards, Lacerta monticola.

J Chem Ecol 32(2):473–488

Lopez P, Ortega J, Martın J (2014) Chemosensory prey detection by

the Amphisbaenian Trogonophis wiegmanni. J Herpetol

48:514–517. doi:10.1670/12-268

Mason RT (1992) Reptilian pheromones. In: Gans C, Crews D (eds)

Biology of the reptilia, vol 18. University of Chicago Press,

Chicago, pp 114–228

Mason RT, Parker MR (2010) Social behavior and pheromonal

communication in reptiles. J Comp Physiol A Neuroethol Sens

Neural Behav Physiol 196(10):729–749

Mehta RS (2008) Early experience shapes the development of

behavioral repertoires of hatchling snakes. J Ethol 27(1):143–151

Mori A (1994) Prey-handling behavior of newly hatched snakes in

two species of the genus Elaphe with comparison to adult

behavior. Ethology 97:198–214

Parker MR, Kardong KV (2005) Rattlesnakes can use airborne cues

during post-strike prey relocation. In: Mason R, LeMaster MP,

Mueller-Schwarze D (eds) Chemical signals in vertebrates 10.

Springer Press, New York, pp 397–402

Persson L, Andersson J, Wahlstrom E, Eklov P (1996) Size specific

interactions in lake systems: predator gape limitation and prey

growth rate and mortality. Ecology 77:900–911

Radcliffe C, Chiszar D, O’Connell B (1980) Effects of prey size on

poststrike behavior in rattlesnakes (Crotalus durissus, C. enyo,

and C. viridis). Bull Psychon Soc 16:449–450

Reformato LS, Kirschenbaum DM, Halpern M (1983) Preliminary

characterization of response-eliciting components of earthworm

extract. Pharmacol Biochem Behav 18:247–254

Rodrıguez-Robles J (2002) Feeding ecology of North American

gopher snakes (Pituophis catenifer, Colubridae). Biol J Linn Soc

77:165–183

Saviola A, Chiszar D, Mackessy S (2012) Ontogenetic shift in

response to prey derived chemical cues in prairie rattlesnakes

Crotalus viridis viridis. Curr Zool 58:549–555

Schubert SN, Houck LD, Feldhoff PW, Feldhoff RC, Woodley SK

(2008) The effects of sex on chemosensory communication in a

terrestrial salamander (Plethodon shermani). Horm Behav

54(2):270–277

Schwenk K (1993) The evolution of chemoreception in squamate

reptiles: a phylogenetic approach. Brain Behav Evol 41:124–137

Scudder K, Stewart N, Smith H (1980) Response of neonate water

snakes (Nerodia s. sipedon) to conspecific chemical cues.

J Herpetol 14(2):196–198

Shepard D, Phillips C, Dreslik M, Jellen B (2004) Prey preference and

diet of neonate eastern massasaugas (Sistrurus c. catenatus). Am

Midl 152:360–368

Shine R, Mason RT (2012) An airborne sex pheromone in snakes.

Biol Lett 8:183–185

Behavioral variation in prey odor responses in northern pine snake neonates and adults

123

Slip DJ, Shine R (1988) Habitat use, movements, and activity patterns

of free-ranging diamond pythons, Morelia s. spilota (Serpentes:

Boidae): a radiotelemetric study. Aust Wildl Res l5:515–553

Stark C, Tiernan C, Chiszar D (2011) Effects of deprivation of

vomeronasal chemoreception on prey discrimination in rat-

tlesnakes. Psychol Rec 61:363–370

Terrick T, Mumme R, Burghardt G (1995) Aposematic coloration

enhances chemosensory recognition of noxious prey in the garter

snake Thamnophis radix. Anim Behav 49:857–866

Troost T, Kooi BW, Dieckmann U (2008) Joint evolution of predator

body size and prey-size preference. Evol Ecol 22:771–799

Urban M (2007) The growth-predation risk trade-off under a growing

gape-limited predation threat. Ecology 88:2587–2597

Urban M (2008) Salamander evolution across a latitudinal cline in

gape limited predation risk. Oikos 117:1037–1049

Waters RM (1993) Odorized air current trailing by garter snakes,

Thamnophis sirtalis. Brain Behav Evol 41:219–223

Waters RM, Burghardt GM (2013) Prey availability influences the

ontogeny and timing of chemoreception-based prey shifting in

the striped crayfish snake, Regina alleni. J Comp Psychol

127(1):49–55

Weldon P, Schell F (1984) Responses by king snakes (Lampropeltis

getulus) to chemicals from colubrid and crotaline snakes. J Chem

Ecol 10:1509–1520

K. P. W. Smith et al.

123